- Email: [email protected]
Ionic liquids monolithic columns for protein separation in capillary electrochromatography
Analytica Chimica Acta 804 (2013) 313–320
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Ionic liquids monolithic columns for protein separation in capillary electrochromatography Cui-Cui Liu, Qi-Liang Deng, Guo-Zhen Fang, Hui-Lin Liu, Jian-Hua Wu, Ming-Fei Pan, Shuo Wang ∗ Key Laboratory of Food Nutrition and Safety, Ministry of Education, Tianjin Key Laboratory of Food Nutrition and Safety, Tianjin University of Science and Technology, Tianjin, 300457, China
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• ILs-monolithic columns with different anions were prepared by two approaches. • The performances of the resulting columns could be designed by tuning anions. • ViOcIm+ NTf2 − based column exhibited the highest column efficiencies for proteins. • ILs gave the columns potential to separate small molecules and macro biomolecules.
a r t i c l e
i n f o
Article history: Received 15 July 2013 Received in revised form 11 October 2013 Accepted 17 October 2013 Available online 25 October 2013 Keywords: Anion effect Capillary electrochromatography Ionic liquid Monolithic column Protein separation
a b s t r a c t A series of ionic liquids (ILs) monolithic capillary columns based on 1-vinyl-3-octylimidazolium (ViOcIm+ ) were prepared by two approaches (“one-pot” approach and “anion-exchange” approach). The effects of different anions (bromide, Br− ; tetrafluoroborate, BF4 − ; hexafluorophosphate, PF6 − ; and bis-trifluoromethanesulfonylimide, NTf2 − ) on chromatography performance of all the resulting columns were investigated systematically under capillary electrochromatography (CEC) mode. The results indicated that all these columns could generate a stable reversed electroosmotic flow (EOF) over a wide pH range from 2.0 to 12.0. For the columns prepared by “one-pot” approach, the EOF decreased in the order of ViOcIm+ Br− > ViOcIm+ BF4 − > ViOcIm+ PF6 − > ViOcIm+ NTf2 − under the same CEC conditions; the ViOcIm+ Br− based column exhibited highest column efficiencies for the test small molecules; the ViOcIm+ NTf2 − based column possessed the strongest retention for aromatic hydrocarbons; and baseline separation of four standard proteins was achieved on ViOcIm+ NTf2 − based column corresponding to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c). These results indicated that the property of ILs based columns could be tuned successfully by changing anions, which gave these columns potential to separate both small molecules and macro biomolecules. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.
Abbreviations: ILs, ionic liquids; RTILs, room temperature ionic liquids; PIL, polymeric ionic liquid; ViOcIm+ Br− , 1-vinyl-3-octylimidazolium bromide; ViOcIm+ BF4 − , 1-vinyl-3-octylimidazolium tetrafluoroborate; ViOcIm+ PF6 − , 1-vinyl-3-octylimidazolium hexafluorophosphate; ViOcIm+ NTf2 − , 1-vinyl-3-octylimidazolium bis-trifluoromethanesulfonylimide; CE, capillary electrophoresis; CEC, capillary electrochromatography; EOF, electroosmotic flow; GC, gas chromatography; HPLC, high performance liquid chromatography; SEM, scanning electron microscopy. ∗ Corresponding author. Tel.: +86 22 6060 1456; fax: +86 22 6060 1332. E-mail addresses: [email protected], [email protected] (S. Wang). 0003-2670/$ – see front matter. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2013.10.037
314
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
1. Introduction Capillary electrochromatography (CEC) which combines superiorities of the capillary electrophoresis (CE) and high performance liquid chromatography (HPLC) has been widely applied in the analytical separation sciences. As the core of CEC, capillary columns have been stimulated to develop by the growing demands for highthroughput analysis and microscale chromatographic separations. Monolithic columns, a novel separation media, have attracted great interests due to the merits of great permeability, fast mass transfer rate, high loading capacity, and ease of preparation [1]. To date, there have been numerous reports on organic polymer-based, inorganic silica-based and organic silica hybrid monolithic columns [2–5]. Organic monolithic columns prepared by in situ polymerization technology exhibits several significant advantages over the silica-based monolith, such as simpler and faster preparation, greater choices of surface functionalities, wider pH stability, as well as better biocompatibility [6]. And their applications in separation have been extended from small molecule to macro biomolecules, especially protein mixture. Nevertheless, they are suffered from apparent disadvantages, such as limited interactions between analytes and the monolithic matrix, high organic solvent content in running buffer and low separation efficiency for protein separation and so on [7–9]. Therefore, new alternatives are highly desirable to overcome these issues. Ionic liquids (ILs) which compose of large asymmetric organic cations and inorganic or organic anions are salts with relatively low melting points compared to most traditional inorganic salts; some of them are even liquid around ambient temperature, called room temperature ionic liquids (RTILs). ILs have recently attracted great interest due to their unique properties such as chemical and thermal stability, non-flammability, non-detectable vapor pressures and chemical tunabilities [10]. Especially, ILs based on imidazolium cations have been proven to be versatile on account of their favorable characteristics, e.g., moisture- and air-stability, easy recycling and being a good solvent for a wider variety of organic and inorganic chemical compounds [11,12]. In addition, imidazolium are “designable” because structural modifications in both the cation (the 1- and 3-positions of the imidazolium ring) and anion permit the tuning of properties [13]. Over the years, ILs have been applied in the different fields of chemistry, such as organic chemistry, inorganic chemistry, electrochemistry, analytical chemistry, and so on [14–19]. The growing interest of ILs in chromatography field can be observed from the dramatic increase in the number of publications appeared during the last decade [20–26]. Especially in CEC field, ILs have usually been used as dynamic capillary coatings, physically adsorbed coatings and covalently linked coatings with the purpose of removing the deleterious effect of free silanols on the retention of basic analytes [27–30]. Recently, polymeric ionic liquid (PIL) has been used as the physically adsorbed coatings in CE for protein separation [29]. Although it is effective to obtain a better stability and a wider pH application range comparing to the traditional coatings (e.g. hydroxyethylcellulose, poly (ethylene oxide) and poly (ethylene glycol) methyl ether methacrylate) [31–33], the repeating wall modification steps between subsequent runs is time and labor consuming. In 2011, a new application of ILs in CEC has been reported by our group. Wang et al. has prepared a novel IL-monolithic capillary column via the thermal free radical copolymerization with IL (1-vinyl-3-octylimidazolium chloride, ViOcIm+ Cl− ), lauryl methacrylate (LMA) and ethylene dimethacrylate (EDMA) in 1,4butanediol/methanol porogen system [34]. As “one-pot” approach, the preparation process dispensing with any additional modification steps is time and labor saving. The resulting column which could generate a stable reversed EOF in a wide pH range (2.0–12.0) not only effectively eliminates the wall adsorption of the basic
analytes but also exhibits great separation efficiency and reproducibility. Considering these superior qualities, we have reasons to expect that ILs-monolithic capillary columns will continue its contribution to protein separation. In this study, a twofold study is reported: (1) 1vinyl-3-octylimidazolium (ViOcIm+ ) based ILs-monolithic capillary columns with different counter ions (bromide, Br− ; tetrafluoroborate, BF4 − ; hexafluorophosphate, PF6 − ; and bistrifluoromethanesulfonylimide, NTf2 − ) were prepared as Wang et al. [34]. (2) Firstly, ViOcIm+ Br− based monolithic capillary columns were prepared as above. Subsequently, anion-exchange was carried out by pumping salt solutions of the anion of interest through the columns for some time. Then all the columns obtained were evaluated chromatographically and applied in protein separation to investigate the impact of anions on separation performance in CEC. 2. Experimental 2.1. Instrumentation CEC experiments were performed on a P/ACE MDQ CE system (Beckman-Coulter, USA) equipped with a UV detector. Data acquisition and processing were controlled by Beckman Chem Station software. Scanning electron microscopy (SEM) micrographs of the monoliths were obtained on a SU1510 SEM (Hitachi, Japan). The elemental (C, H, O, N) contents of the prepared monoliths were determined on Vario MACRO cube (ELEMENTAR, Germany) by using TCD detector. 2.2. Chemicals and materials 1-Vinylimidazole, LMA, ␥-methacryloxypropyltrimethoxysilane (␥-MAPS) and ethylene dimethacrylate (EDMA) were obtained from Sigma (St. Louis, MO, USA). 1-Bromooctane was purchased from TCI (Tokyo, Japan). The free radical initiator 2,2-azobisisobutyronitrile (AIBN, 99%) was obtained from Tianjin Chemical Reagent Factory (Tianjin, China) and recrystallized in ethanol before use. Sodium tetrafluoroborate (NaBF4 ), potassium hexafluorophosphate (KPF6 ) and lithium bis-trifluoromethanesulfonylimide (LiNTf2 ) were purchased from Fluka (Fluka, Buchs, Switzerland). A fused-silica capillary (100 m i.d., 375 m o.d.) was purchased from the Yongnian Optic Fiber Plant (Hebei, China). Doubly deionized water (DDW, 18 M cm−1 ) produced using a Milli-Q system (Millipore Corporation, USA) was used throughout the experiments. Chromatographic grade of acetonitrile (ACN) were purchased from Tianjin Chemical Plant (Tianjin, China), and the other chemicals were at least of analytical grade. 2.3. Synthesis of ILs ViOcIm+ Br− was synthesized according to the protocol reported by Hsieh et al. [35]. Briefly, 1-bromooctane (9.270 g, 0.048 mol) was added dropwise to 1-vinylimidazole (3.760 g, 0.040 mol). The mixture was heated to 70 ◦ C under stirring for 50 h. Phase separation occurred and the viscous brown liquid obtained was washed with ethyl acetate. Then the product was filtered and dried in a vacuum oven until constant weight. Different counter ion ILs were prepared by modifying the procedures described in the literature [36]. Simple anion-exchange reactions were conducted to replace the bromide ions. 2.880 g of ViOcIm+ Br− was dissolved in 10 mL of distilled water, and 1.096 g of NaBF4 in 10 mL water was slowly added. After stirring for 12 h at room temperature, the resulting viscous brown liquid was washed thoroughly with distilled water and dried in a vacuum oven until
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
constant weight. As for the ViOcIm+ PF6 − and ViOcIm+ NTf2 − , KPF6 and LiNTf2 were used as the anion-exchange reagents, respectively. 2.4. Preparation of ILs-monolithic capillary columns A bare capillary was preconditioned prior to its first use by flushing with 1 M NaOH for 3 h, H2 O for 30 min, 1 M HCl for 1 h, H2 O for another 30 min and methanol for 30 min in sequence. The capillary inner wall was then allowed to react with a solution of 50% ␥-MAPS in methanol in a 40 ◦ C water bath overnight to vinylize the inner wall of the capillary. Finally, the capillary was rinsed with methanol to flush out the residuals and then dried with a stream of nitrogen for further use. The brief process for preparing ILs-monolithic capillary columns is as follow. Approach 1: The polymerization mixture consisting of 0.170 g ViOcIm+ Br− (0.174 g ViOcIm+ BF4 − , 0.209 g ViOcIm+ PF6 − or 0.289 g ViOcIm+ NTf2 − ), 0.096 g LMA, 0.160 g EDMA, 0.004 g AIBN, 0.600 g acetone and 0.600 g 1,4-butanediol were sonicated for 15 min to degas. The obtained homogeneous solutions were manually injected into the ␥-MAPS pretreated capillaries to an appropriate length by a syringe. After both ends were sealed with silicone rubbers, the capillaries were incubated in a 60 ◦ C water bath for 24 h. The resulting monolithic columns were washed with methanol by a HPLC pump to remove unreacted monomers and porogen. As detection window was made by burning off a 2 mm segment of the protecting polymer layer at the end of the monolithic bed, the monolithic columns whose total length was 32 cm with an effective length of 20 cm were tested chromatographically. Approach 2: ViOcIm+ Br− based monolithic columns were prepared as approach 1, and then anion-exchange was carried out in the columns by pumping 0.5 M salt solutions of the anion of interest (NaBF4 , KPF6 or LiNTf2 ) at 5 L min−1 until the retention time of toluene on the resulting columns was constant. Then the anion exchange ratio was tested by ICP.
copolymerization happened among ␥-MAPS, ViOcIm+ Br− , LMA and EDMA. Moreover, it was obvious that the matrix structure of ViOcIm+ Br− based column was much denser than the other three columns, especially ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns. As can be seen from Fig. 1b, d, f and h that the particle size of monolithic matrix increase progressively in the order ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − . of Brunauer–Emmett–Teller (BET) method was used to determine average pore diameter as 15.53 nm, 22.49 nm, 25.36 nm and 31.08 nm for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns, respectively. And the specific surface area of the monoliths was found to be 9.29 m2 g−1 , 7.95 m2 g−1 , 6.70 m2 g−1 and 6.57 m2 g−1 for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns, respectively. These results indicated that different anions had significant influence on preparation process of monolithic columns. The C%, H%, N% and O% (w/w) of these monolithic columns were determined by element analysis. As shown in Table 1, the contents of C and H immobilized on the monolithic matrix decreased in the order of ViOcIm+ Br− > ViOcIm+ BF4 − > ViOcIm+ PF6 − > ViOcIm+ NTf2 − . The content of N also decreased in the same order except for ViOcIm+ NTf2 − based column (whose anion contains N element). The cause of changes in elements content might result from the increased molecular weight of anions. Additionally, for the same anion based columns, N content of column prepared by approach 1 was lower than that of column prepared by approach 2. This may be attributed to the effect of the anions on the copolymerization process. The B%, P% and S% (w/w) of the columns prepared by approach 2 were determined as 0.63% for B, 1.75% for P and 2.14% for S. According to these results, the anion exchange ratio of the columns prepared by approach 2 could be calculated as 89.23% for BF4 − based column, 89.69% for PF6 − based column and 84.21% for NTf2 − based column. The permeability of the columns was evaluated by the Darcy’s Law [37]:
2.5. CEC procedures B0 = The resulting monolithic columns were first conditioned by running buffer for about 30 min with a manual syringe pump, and then equilibrated at a low voltage (−5 kV, ramping time for 10 min) until a stable current was obtained. Separations were performed at 25 ◦ C using different voltages and running buffer (stated in the figure captions). Stock solutions of proteins, thiourea and its analogs were prepared with distilled water (2 mg mL−1 for proteins, 0.05 mg mL−1 for thiourea and its analogs). Standard solutions of alkyl benzenes were diluted to an appropriate concentration with ethanol. All the test mixture were prepared by mixing the standard stock solutions, diluted with buffer in an appropriate ratio and stored in the refrigerator. The egg was purchased locally and egg white was diluted with 10 mM PBS (pH 7.4) in a 1:20 ratio [31]. The mixture was oscillated for 1 min and then centrifuged at 600 × g for 15 min. Prior to use, all solutions and samples were ultrasonically degassed and filtered through a 0.45 m syringe filter.
315
FL r 2 P
(1)
where F is the flow rate of the mobile phase (m3 s−1 ), is the viscosity of the mobile phase (Pa s), L is the effective length of column (m), r is the inner radius of the column (m), and P is the pressure drop of the column (Pa). The B0 value of the four monolithic columns (from ViOcIm+ Br− to ViOcIm+ NTf2 − ) was calculated as 2.05, 2.70, 3.32, 4.02 × 10−13 m2 for ACN ( = 0.38 cP), which indicated the good permeability of the resulting columns. In addition, all the columns promised excellent mechanical stability without movement of the monolithic bed under a pressure of up to 30 MPa for at least 48 h. 3.2. EOF of ILs-monolithic capillary columns Since EOF is the basic requirement for driving mobile phases through a capillary column, it is an effective indicator to the evaluation of the monolithic columns calculated as the following equations: Le × Lt V × t0
3. Results and discussion
EOF =
3.1. Characterization of ILs-monolithic capillary columns
where Le , Lt , V, t0 are the effective length (cm), total length of column (cm), the applied voltage (V), the retention time of thiourea (s), respectively. Here, ILs acted as the main source of the EOF from cathode to anode. To investigate the effects of different anions on EOF, the relationship between the pH of the running buffer and the EOF of ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns was examined. The results indicated
The cross-section morphology of the four monolithic columns prepared by approach 1 was obtained by SEM. In Fig. 1, it could be seen that the formed monolithic matrix was attached to the inner capillary wall well. This was because of not only the successful pretreatment of capillary by ␥-MAPS but also the perfect
(2)
316
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 1. SEM images of the ILs-monolithic capillary columns prepared by approach 1 (ViOcIm+ Br− based column: (a) 600×; (b) 5000×; ViOcIm+ BF4 − based column: (c) 600×; (d) 5000×; ViOcIm+ PF6 − based column: (e) 600×; (f) 5000×; ViOcIm+ NTf2 − based column: (g) 600×; (h) 5000×).
Table 1 The C%, H%, N% and O% (w/w) of the columns. C (%)
H (%)
N (%)
O (%)
Column 1 Br− BF4 − PF6 − NTf2 −
63.36 63.34 62.36 61.29
8.62 8.61 8.38 8.22
1.84 1.60 1.23 1.40
21.01 21.48 21.65 23.19
Column 2 BF4 − PF6 − NTf2 −
63.10 61.03 58.24
8.59 8.30 7.76
1.83 1.77 2.35
20.92 20.23 22.09
that all the columns could generate strong reversed EOF (greater than 1.60 × 10−4 cm2 V−1 s−1 ) even at a high pH 12, which indicated the successful copolymerization of ILs onto the columns. Furthermore, as shown in Fig. 2, EOF of columns prepared by approach 1 was reduced obviously from ViOcIm+ Br− to ViOcIm+ NTf2 − based column at the same test pH. It might be on account of the decrease in charge number on monolithic matrix, which was just consistent with our previous supposition that the amount of imidazolium ring on the monolithic matrix decreased from ViOcIm+ Br− to ViOcIm+ NTf2 − . Note that there was almost no visible difference in EOF among the four columns prepared by approach 2, which was almost equal to EOF of ViOcIm+ Br− based column (data not shown).
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 2. Relationship between the EOF of the ILs-monolithic capillary columns prepared by approach 1 and the pH of the running buffer. CEC conditions: 30 mM phosphate buffer (H3 PO4 -Na2 HPO4 for pH 2.0–9.0, Na2 HPO4 –NaOH for 9.0–12.0) containing 30% (v/v) ACN with various pH (pH 2.0–12.0); separation voltage, − 10 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm; EOF marker, thiourea.
317
Phenol and aniline compounds were used to further demonstrate the application of these ILs-columns. Fig. 4 showed these polar compounds were separated with elution order of phenol < aniline < m-nitrophenol < o-nitroaniline < hydroquinone, and the retention time of these analytes on the columns was prolonged in the order of ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − . Additionally, for the columns prepared by approach 1, the resolution was reduced from 1.97 to 0.75 for m-nitrophenol and o-nitroaniline in the order ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − of (Fig. 4A). While, for the columns prepared by approach 2, the resolution was improved from 1.97 to 2.28 for m-nitrophenol and o-nitroaniline in the same order, and the highest separation efficiency was obtained on ViOcIm+ NTf2 − based column corresponding to column efficiency of 28 000–58 000 N m−1 (Fig. 4B). These results showed that different anions based columns had different effects on the separation of these phenol and aniline compounds, and the same anion based columns prepared by two approaches also exhibited different separation performance for these analytes. Additionally, the hydrophobicity of the long chain as well as the electropositivity and aromaticity of the imidazole on the stationary phase might cooperatively participate in the separation of phenol and aniline compounds. 3.4. Separation of thiourea and its analogs
3.3. Separation of aromatic compounds The monolithic columns prepared by two approaches were tested chromatographically to study the effects of anions on the separation performance. As shown in Fig. 3, four alkylbenzenes could be baseline separated on each column with the eluting sequence of thiourea < toluene < ethylbenzene < propylbenzene < butylbenzene. Also, the retention time for alkylbenzenes was prolonged with hydrophobicity of the anions increasing (Br− < BF4 − < PF6 − < NTf2 − ). These results indicated a reversed-phase separation of these compounds on the columns. Note that, for the same anion based columns, the retention of alkylbenzenes on columns prepared by approach 1 was stronger than that on columns prepared by approach 2. It might be ascribed to the increase in density of hydrophobic functional groups on the monolithic matrix, which was calculated based on the ratio of the content of C and H to the specific surface area of the monoliths as 77.48 mg m−2 , 90.50 mg m−2 , 105.58 mg m−2 , 105.80 mg m−2 for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − , ViOcIm+ NTf2 − based column prepared by approach 1, respectively. As to the columns prepared by approach 2, the density of hydrophobic functional groups could not be improved by anion-exchange.
In this study, the effects of different anions based columns on the separation of analogs were also investigated. The results showed that three analogs could be baseline separated on ViOcIm+ Br− based column with the elution order dimethyl sulfoxide < thiourea < N-methylthiourea, which was consistent with the hydrophobicity of the analytes (Fig. 5A). Nevertheless, they could be partially separated on the ViOcIm+ BF4 − and ViOcIm+ PF6 − based columns and almost co-eluted on the ViOcIm+ NTf2 − based column prepared by approach 1. It might be ascribed to the different matrix structures and average pore diameters of the resulting columns based on different anions. In order to verify further, the separation of thiourea and its analogs was also attempted on columns prepared by approach 2 (Fig. 5B). The results showed that the four columns exhibited prolonged retention time and improved separation efficiency (from 31 000 to 34 000 N m−1 for dimethyl sulfoxide) with hydrophobicity of the anions increasing (Br− < BF4 − < PF6 − < NTf2 − ). These results illustrated that the hydrophobicity of the long chain and the matrix structures resulted from the anions owning different characteristics (such as size, viscosity, polarity and so on) [38–40] corporately played a role in the separation of thiourea and its analogs.
Fig. 3. Separation of alkylbenzenes on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM acetic acid buffer containing 40% (v/v) ACN at pH 3.0. Other CEC conditions were same as in Fig. 2. Analytes: 1, thiourea; 2, toluene; 3, ethylbenzene; 4, propylbenzene; 5, butylbenzene.
318
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 4. Separation of phenol and aniline compounds on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM NaH2 PO4 buffer containing 30% (v/v) at pH 4.0; separation voltage, − 10 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm. Analytes: 1, phenol; 2, aniline; 3, m-nitrophenol; 4, o-nitroaniline; 5, hydroquinone.
Fig. 5. Separation of thiourea and its analogs on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM H3 PO4 –Na2 HPO4 buffer at pH 4.0; separation voltage, − 2 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm. Analytes: 1, dimethyl sulfoxide; 2, thiourea; 3, N-methylthiourea.
3.5. Protein separation Here, the potential ability of all the resulting monolithic columns with different anions for the separation of protein mixture was demonstrated by four standard proteins including bovine serum albumin (BSA, molecular size 5.0 × 7.0 × 7.0 nm, isoelectric point (pI) 4.8), equine myoglobin (Mb, molecular size 4.5 × 3.5 × 2.5 nm, pI 7.0), lysozyme (Lyz, molecular size 3.0 × 3.0 × 4.5 nm, pI 11.0) and cytochrome c (Cyt c, molecular size 2.6 × 3.2 × 3.0 nm, pI 10.7). As shown in Fig. 6, they could be baseline separated within 12 min on the ViOcIm+ NTf2 − based column prepared by approach 1 using 20% (v/v) ACN, 30 mM Na2 HPO4 –citric acid buffer as background electrolyte with the elution order of Cyt c < Lyz < Mb < BSA. The separation efficiency was improved greatly compared to previous report [31], which corresponded to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c). The monoliths prepared by approach 2 exhibited poor resolution (Rs < 1) for these proteins under the same separation conditions although the elution order of the four proteins accorded with that on monoliths prepared by approach 1 (data not shown). It might be attributed to the effects of different anions on formation of matrix structures. To study separation mechanism further, the effect of ACN content (v/v) in buffer on retention factors (k) of the proteins was investigated on ViOcIm+ NTf2 − based column prepared by approach 1. The k decreased with the ACN content increasing from 15% to 40%. In our experiments, with less than 20% ACN in buffer, the great k would result in lower column efficiencies of proteins. When ACN content in buffer was over 30% the protein mixture would
be eluted together. So 20% ACN in buffer was chosen for protein separation. These results concluded that the hydrophobic interaction contributed to protein separation. Here, the organic solvent content in buffer was greatly reduced compared to the previous
Fig. 6. Separation of proteins on the ILs-monolithic capillary columns prepared by approach 1. (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 20% (v/v) ACN, 30 mM Na2 HPO4 -citric acid buffer at pH 3.0; separation voltage, −10 kV; injection, 5 psi for 10 s; detection wavelength, 210 nm. Analytes: 1, cytochrome c; 2, lysozyme; 3, equine myoglobin; 4, bovine serum albumin.
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
319
To further validate the potential of the ViOcIm+ NTf2 − based column prepared by approach 1, it was also evaluated for the separation of the egg white sample. As shown in Fig. 8, 8 proteins were positively identified, demonstrating the enhanced separation performance of our column compared with the previously report [31]. 3.6. Reproducibility of ILs-monolithic capillary columns The reproducibility of the monolithic columns prepared by approach 1 was evaluated by measuring the relative standard deviations (RSDs) of the retention time and the peak area for four alkylbenzenes. The RSDs of the retention time and the peak area for intra-column assay were in the range of 0.88–1.42% and 3.06–5.56% (n = 5), respectively. Inter-column reproducibility for the retention time and the peak area was also evaluated which were in the range of 2.90–4.86 and 4.73–7.26%, respectively. These results indicated that the present columns possessed a robust and repeatable separation performance. Fig. 7. Effect of pH for running buffer on separation of proteins on ViOcIm+ NTf2 − based column prepared by approach 1. CEC conditions: 20% (v/v) ACN, 30 mM Na2 HPO4 –citric acid buffer at different pH. Other conditions and peak identification were as in Fig. 6.
report [9,41], which was worthy mentioning in protein separation. In addition, according to the elution order, the separation of the four proteins was consistent with ion exclusion mechanism [42] where similarly charged species to the phase are repelled, while oppositely charged and neutral species are retained, since both the imidazolium rings on stationary phase and proteins were positively charged under the test CEC conditions. The buffer pH strongly affects the protein separation because it can remarkably change the net charges. Here, the effect of pH on separation of the four standard proteins was also evaluated on ViOcIm+ NTf2 − based column prepared by approach 1. As can be seen from Fig. 7 that the retention of proteins increased, and the peaks became wider with pH increasing from 2.5 to 4.0. It might be attributed to the decrease in EOF. Besides, the decrease in the net positive charges of proteins resulted in a decrease of the repulsion forces between proteins and the imidazolium ring, which could also cause retention to increase and protein adsorption, especially BSA (pI 4.8).
4. Conclusions In this study, a series of ViOcIm+ based ILs-monolithic columns with different anions (Br− , BF4 − , PF6 − and NTf2 − ) were prepared by two approaches (“one-pot” approach and “anion-exchange” approach). The results indicated that morphology and chromatography performance of the resulting columns could be tuned by changing the kind of anions. Under CEC mode, the ViOcIm+ Br− based column exhibited highest column efficiencies for the test small molecules. The ViOcIm+ NTf2 − based column possessed the strongest retention for aromatic hydrocarbons. Their potential ability for the separation of protein mixture was demonstrated by four standard proteins, which corresponded to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c) on ViOcIm+ NTf2 − based column. To the best of our knowledge, this is the first time to examine and discuss the anion effects on the separation performance of monolithic columns. It means that ILs expand the variability of stationary phases of chromatography, and they are of great interest in separation science. Acknowledgements The authors are grateful to financial support from the Ministry of Science and Technology of China (Project No. 2012AA101609), the National Natural Science Foundation of China (21075089), the Ministry of Science and Technology of Tianjin (Project No. 10SYSYJC28300) and the Program for Changjiang Scholars and Innovative Research Team in University (Project No. IRT1166). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
+
−
Fig. 8. Separation of the egg white sample on ViOcIm NTf2 based column prepared by approach 1. CEC conditions: mobile phase, 40 mM Na2 HPO4 –citric acid buffer at pH 4.0; applied voltage, −2 kV; injection, 5.0 psi for 10 s; detection wavelength, 210 nm. Analytes: the egg white sample.
[12] [13]
G. Guiochon, J. Chromatogr. A 1168 (2007) 101–168. J.X. He, G.Z. Fang, Q.L. Deng, S. Wang, Anal. Chim. Acta 692 (2011) 57–62. J. Rozenbrand, W.P. van Bennekom, J. Sep. Sci. 34 (2011) 1934–1944. J.N. Hay, H.M. Raval, Chem. Mater. 13 (2011) 3396–3403. S.D. Chambers, F. Svec, J.M.J. Frechet, J. Chromatogr. A 1218 (2011) 2546–2552. Z. Lin, H. Huang, X. Sun, Y. Lin, L. Zhang, G. Chen, J. Chromatogr. A 1246 (2012) 90–97. J. Liu, L. Ren, Y. Liu, H. Li, Z. Liu, J. Chromatogr. A 1228 (2012) 276–282. M. Iwasaki, N. Sugiyama, N. Tanaka, Y. Ishihama, J. Chromatogr. A 1228 (2012) 292–297. D. Bandilla, C.D. Skinner, J. Chromatogr. A 1004 (2003) 167–179. H.S. Schrekker, M.P. Stracke, C.M.L. Schrekker, J. Dupont, Ind. Eng. Chem. Res. 46 (2007) 7389–7392. Z.B. Alfassi, R.E. Huie, B.L. Milman, P. Neta, Anal. Bioanal. Chem. 377 (2003) 159–164. J.L. Anthony, E.J. Maginn, J.F. Brennecke, J. Phys. Chem. B 105 (2001) 10942–10949. X.A. Tian, X.P. Zhang, L. Wei, S.J. Zeng, L. Huang, S.J. Zhang, Green Chem. 12 (2010) 1263–1273.
320 [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320 P. Domínguez de María, Angew. Chem. 120 (2008) 7066–7075. I.J.B. Lin, C.S. Vasam, J. Organomet. Chem. 690 (2005) 3498–3512. L.H. Lu, W. Chen, Adv. Mater. 22 (2010) 3745–3748. C.M.S.S. Neves, J.F.O. Granjo, M.G. Freire, A. Robertson, N.M.C. Oliveira, J.A.P. Coutinho, Green Chem. 13 (2011) 1517–1526. L. Vidal, M.L. Riekkola, A. Canals, Anal. Chim. Acta 715 (2012) 19–41. R. Giernoth, Angew. Chem. Int. Ed. 49 (2010) 2834–2839. T. Payagala, Y. Zhang, E. Wanigasekara, K. Huang, Z.S. Breitbach, P.S. Sharma, L.M. Sidisky, D.W. Armstrong, Anal. Chem. 81 (2009) 160–173. T.T. Ho, C.Y. Chen, Z.G. Li, T.C.C. Yang, M.R. Lee, Anal. Chim. Acta 712 (2012) 72–77. W.C. Siegler, J.A. Crank, D.W. Armstrong, R.E. Synovec, J. Chromatogr. A 1217 (2010) 3144–3149. C.F. Poole, S.K. Poole, J. Sep. Sci. 34 (2011) 888–900. J.L. Anderson, D.W. Armstrong, Anal. Chem. 77 (2005) 6453–6462. A. Martin-Calero, V. Pino, J.H. Ayala, V. González, A.M. Afonso, Talanta 79 (2009) 590–597. M. Bian, Z.J. Zhang, H. Yin, J. Pharm. Biomed. Anal. 58 (2012) 163–167. Y. Tian, R. Feng, L.P. Liao, H.L. Liu, H. Chen, Z.R. Zeng, Electrophoresis 29 (2008) 3153–3159. H.F. Han, J. Li, X.S. Wang, X. Liu, S.X. Jiang, J. Sep. Sci. 34 (2011) 2323–2328.
[29] J. Li, H.F. Han, Q. Wang, X. Liu, S.X. Jiang, J. Sep. Sci. 34 (2011) 1555–1560. [30] H. Han, Q. Wang, X. Liu, S. Jiang, J. Chromatogr. A 1246 (2012) 9–14. [31] D. Zhou, L.N. Xiang, R.J. Zeng, F.H. Cao, X.X. Zhu, Y.M. Wang, J. Sep. Sci. 34 (2011) 3441–3450. [32] M.H. Kleemiss, M. Gilges, G. Schomburg, Electrophoresis 14 (1993) 515–522. [33] C.A. Lucy, A.M. MacDonald, M.D. Gulcev, J. Chromatogr. A 1184 (2008) 81–105. [34] Y. Wang, Q.L. Deng, G.Z. Fang, M.F. Pan, Y. Yu, S. Wang, Anal. Chim. Acta 712 (2012) 1–8. [35] Y.N. Hsieh, W.Y. Ho, R.S. Horng, P.C. Huang, C.Y. Hsu, H.H. Huang, C.H. Kuei, Chromatographia 66 (2007) 607–611. [36] R. Marcilla, J.A. Blazquez, J. Rodriguez, J.A. Pomposo, D. Mecerreyes, J. Polym. Sci. Part A: Polym. Chem. 42 (2004) 208–212. [37] R.D. Stanelle, L.C. Sander, R.K. Marcus, J. Chromatogr. A 1100 (2005) 68–75. [38] R. Losada, C. Wandrey, Macromolecules 42 (2009) 3285–3293. [39] Y.H. Shih, B. Singco, W.L. Liu, C.H. Hsu, H.Y. Huang, Green Chem. 13 (2011) 296–299. [40] H. Tokuda, K. Hayamizu, K. Ishii, M. Abu Bin, Susan Hasan, M. Watanabe, J. Phys. Chem. B 108 (2004) 16593–16600. [41] M.H. Wu, R.A. Wu, R.B. Li, H.Q. Qin, J. Dong, Z.B. Zhang, H.F. Zou, Anal. Chem. 82 (2010) 5447–5454. [42] P.R. Haddad, M. Noviˇc, J. Chromatogr. A 1118 (2006) 19–28.
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Ionic liquids monolithic columns for protein separation in capillary electrochromatography Cui-Cui Liu, Qi-Liang Deng, Guo-Zhen Fang, Hui-Lin Liu, Jian-Hua Wu, Ming-Fei Pan, Shuo Wang ∗ Key Laboratory of Food Nutrition and Safety, Ministry of Education, Tianjin Key Laboratory of Food Nutrition and Safety, Tianjin University of Science and Technology, Tianjin, 300457, China
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• ILs-monolithic columns with different anions were prepared by two approaches. • The performances of the resulting columns could be designed by tuning anions. • ViOcIm+ NTf2 − based column exhibited the highest column efficiencies for proteins. • ILs gave the columns potential to separate small molecules and macro biomolecules.
a r t i c l e
i n f o
Article history: Received 15 July 2013 Received in revised form 11 October 2013 Accepted 17 October 2013 Available online 25 October 2013 Keywords: Anion effect Capillary electrochromatography Ionic liquid Monolithic column Protein separation
a b s t r a c t A series of ionic liquids (ILs) monolithic capillary columns based on 1-vinyl-3-octylimidazolium (ViOcIm+ ) were prepared by two approaches (“one-pot” approach and “anion-exchange” approach). The effects of different anions (bromide, Br− ; tetrafluoroborate, BF4 − ; hexafluorophosphate, PF6 − ; and bis-trifluoromethanesulfonylimide, NTf2 − ) on chromatography performance of all the resulting columns were investigated systematically under capillary electrochromatography (CEC) mode. The results indicated that all these columns could generate a stable reversed electroosmotic flow (EOF) over a wide pH range from 2.0 to 12.0. For the columns prepared by “one-pot” approach, the EOF decreased in the order of ViOcIm+ Br− > ViOcIm+ BF4 − > ViOcIm+ PF6 − > ViOcIm+ NTf2 − under the same CEC conditions; the ViOcIm+ Br− based column exhibited highest column efficiencies for the test small molecules; the ViOcIm+ NTf2 − based column possessed the strongest retention for aromatic hydrocarbons; and baseline separation of four standard proteins was achieved on ViOcIm+ NTf2 − based column corresponding to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c). These results indicated that the property of ILs based columns could be tuned successfully by changing anions, which gave these columns potential to separate both small molecules and macro biomolecules. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.
Abbreviations: ILs, ionic liquids; RTILs, room temperature ionic liquids; PIL, polymeric ionic liquid; ViOcIm+ Br− , 1-vinyl-3-octylimidazolium bromide; ViOcIm+ BF4 − , 1-vinyl-3-octylimidazolium tetrafluoroborate; ViOcIm+ PF6 − , 1-vinyl-3-octylimidazolium hexafluorophosphate; ViOcIm+ NTf2 − , 1-vinyl-3-octylimidazolium bis-trifluoromethanesulfonylimide; CE, capillary electrophoresis; CEC, capillary electrochromatography; EOF, electroosmotic flow; GC, gas chromatography; HPLC, high performance liquid chromatography; SEM, scanning electron microscopy. ∗ Corresponding author. Tel.: +86 22 6060 1456; fax: +86 22 6060 1332. E-mail addresses: [email protected], [email protected] (S. Wang). 0003-2670/$ – see front matter. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2013.10.037
314
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
1. Introduction Capillary electrochromatography (CEC) which combines superiorities of the capillary electrophoresis (CE) and high performance liquid chromatography (HPLC) has been widely applied in the analytical separation sciences. As the core of CEC, capillary columns have been stimulated to develop by the growing demands for highthroughput analysis and microscale chromatographic separations. Monolithic columns, a novel separation media, have attracted great interests due to the merits of great permeability, fast mass transfer rate, high loading capacity, and ease of preparation [1]. To date, there have been numerous reports on organic polymer-based, inorganic silica-based and organic silica hybrid monolithic columns [2–5]. Organic monolithic columns prepared by in situ polymerization technology exhibits several significant advantages over the silica-based monolith, such as simpler and faster preparation, greater choices of surface functionalities, wider pH stability, as well as better biocompatibility [6]. And their applications in separation have been extended from small molecule to macro biomolecules, especially protein mixture. Nevertheless, they are suffered from apparent disadvantages, such as limited interactions between analytes and the monolithic matrix, high organic solvent content in running buffer and low separation efficiency for protein separation and so on [7–9]. Therefore, new alternatives are highly desirable to overcome these issues. Ionic liquids (ILs) which compose of large asymmetric organic cations and inorganic or organic anions are salts with relatively low melting points compared to most traditional inorganic salts; some of them are even liquid around ambient temperature, called room temperature ionic liquids (RTILs). ILs have recently attracted great interest due to their unique properties such as chemical and thermal stability, non-flammability, non-detectable vapor pressures and chemical tunabilities [10]. Especially, ILs based on imidazolium cations have been proven to be versatile on account of their favorable characteristics, e.g., moisture- and air-stability, easy recycling and being a good solvent for a wider variety of organic and inorganic chemical compounds [11,12]. In addition, imidazolium are “designable” because structural modifications in both the cation (the 1- and 3-positions of the imidazolium ring) and anion permit the tuning of properties [13]. Over the years, ILs have been applied in the different fields of chemistry, such as organic chemistry, inorganic chemistry, electrochemistry, analytical chemistry, and so on [14–19]. The growing interest of ILs in chromatography field can be observed from the dramatic increase in the number of publications appeared during the last decade [20–26]. Especially in CEC field, ILs have usually been used as dynamic capillary coatings, physically adsorbed coatings and covalently linked coatings with the purpose of removing the deleterious effect of free silanols on the retention of basic analytes [27–30]. Recently, polymeric ionic liquid (PIL) has been used as the physically adsorbed coatings in CE for protein separation [29]. Although it is effective to obtain a better stability and a wider pH application range comparing to the traditional coatings (e.g. hydroxyethylcellulose, poly (ethylene oxide) and poly (ethylene glycol) methyl ether methacrylate) [31–33], the repeating wall modification steps between subsequent runs is time and labor consuming. In 2011, a new application of ILs in CEC has been reported by our group. Wang et al. has prepared a novel IL-monolithic capillary column via the thermal free radical copolymerization with IL (1-vinyl-3-octylimidazolium chloride, ViOcIm+ Cl− ), lauryl methacrylate (LMA) and ethylene dimethacrylate (EDMA) in 1,4butanediol/methanol porogen system [34]. As “one-pot” approach, the preparation process dispensing with any additional modification steps is time and labor saving. The resulting column which could generate a stable reversed EOF in a wide pH range (2.0–12.0) not only effectively eliminates the wall adsorption of the basic
analytes but also exhibits great separation efficiency and reproducibility. Considering these superior qualities, we have reasons to expect that ILs-monolithic capillary columns will continue its contribution to protein separation. In this study, a twofold study is reported: (1) 1vinyl-3-octylimidazolium (ViOcIm+ ) based ILs-monolithic capillary columns with different counter ions (bromide, Br− ; tetrafluoroborate, BF4 − ; hexafluorophosphate, PF6 − ; and bistrifluoromethanesulfonylimide, NTf2 − ) were prepared as Wang et al. [34]. (2) Firstly, ViOcIm+ Br− based monolithic capillary columns were prepared as above. Subsequently, anion-exchange was carried out by pumping salt solutions of the anion of interest through the columns for some time. Then all the columns obtained were evaluated chromatographically and applied in protein separation to investigate the impact of anions on separation performance in CEC. 2. Experimental 2.1. Instrumentation CEC experiments were performed on a P/ACE MDQ CE system (Beckman-Coulter, USA) equipped with a UV detector. Data acquisition and processing were controlled by Beckman Chem Station software. Scanning electron microscopy (SEM) micrographs of the monoliths were obtained on a SU1510 SEM (Hitachi, Japan). The elemental (C, H, O, N) contents of the prepared monoliths were determined on Vario MACRO cube (ELEMENTAR, Germany) by using TCD detector. 2.2. Chemicals and materials 1-Vinylimidazole, LMA, ␥-methacryloxypropyltrimethoxysilane (␥-MAPS) and ethylene dimethacrylate (EDMA) were obtained from Sigma (St. Louis, MO, USA). 1-Bromooctane was purchased from TCI (Tokyo, Japan). The free radical initiator 2,2-azobisisobutyronitrile (AIBN, 99%) was obtained from Tianjin Chemical Reagent Factory (Tianjin, China) and recrystallized in ethanol before use. Sodium tetrafluoroborate (NaBF4 ), potassium hexafluorophosphate (KPF6 ) and lithium bis-trifluoromethanesulfonylimide (LiNTf2 ) were purchased from Fluka (Fluka, Buchs, Switzerland). A fused-silica capillary (100 m i.d., 375 m o.d.) was purchased from the Yongnian Optic Fiber Plant (Hebei, China). Doubly deionized water (DDW, 18 M cm−1 ) produced using a Milli-Q system (Millipore Corporation, USA) was used throughout the experiments. Chromatographic grade of acetonitrile (ACN) were purchased from Tianjin Chemical Plant (Tianjin, China), and the other chemicals were at least of analytical grade. 2.3. Synthesis of ILs ViOcIm+ Br− was synthesized according to the protocol reported by Hsieh et al. [35]. Briefly, 1-bromooctane (9.270 g, 0.048 mol) was added dropwise to 1-vinylimidazole (3.760 g, 0.040 mol). The mixture was heated to 70 ◦ C under stirring for 50 h. Phase separation occurred and the viscous brown liquid obtained was washed with ethyl acetate. Then the product was filtered and dried in a vacuum oven until constant weight. Different counter ion ILs were prepared by modifying the procedures described in the literature [36]. Simple anion-exchange reactions were conducted to replace the bromide ions. 2.880 g of ViOcIm+ Br− was dissolved in 10 mL of distilled water, and 1.096 g of NaBF4 in 10 mL water was slowly added. After stirring for 12 h at room temperature, the resulting viscous brown liquid was washed thoroughly with distilled water and dried in a vacuum oven until
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
constant weight. As for the ViOcIm+ PF6 − and ViOcIm+ NTf2 − , KPF6 and LiNTf2 were used as the anion-exchange reagents, respectively. 2.4. Preparation of ILs-monolithic capillary columns A bare capillary was preconditioned prior to its first use by flushing with 1 M NaOH for 3 h, H2 O for 30 min, 1 M HCl for 1 h, H2 O for another 30 min and methanol for 30 min in sequence. The capillary inner wall was then allowed to react with a solution of 50% ␥-MAPS in methanol in a 40 ◦ C water bath overnight to vinylize the inner wall of the capillary. Finally, the capillary was rinsed with methanol to flush out the residuals and then dried with a stream of nitrogen for further use. The brief process for preparing ILs-monolithic capillary columns is as follow. Approach 1: The polymerization mixture consisting of 0.170 g ViOcIm+ Br− (0.174 g ViOcIm+ BF4 − , 0.209 g ViOcIm+ PF6 − or 0.289 g ViOcIm+ NTf2 − ), 0.096 g LMA, 0.160 g EDMA, 0.004 g AIBN, 0.600 g acetone and 0.600 g 1,4-butanediol were sonicated for 15 min to degas. The obtained homogeneous solutions were manually injected into the ␥-MAPS pretreated capillaries to an appropriate length by a syringe. After both ends were sealed with silicone rubbers, the capillaries were incubated in a 60 ◦ C water bath for 24 h. The resulting monolithic columns were washed with methanol by a HPLC pump to remove unreacted monomers and porogen. As detection window was made by burning off a 2 mm segment of the protecting polymer layer at the end of the monolithic bed, the monolithic columns whose total length was 32 cm with an effective length of 20 cm were tested chromatographically. Approach 2: ViOcIm+ Br− based monolithic columns were prepared as approach 1, and then anion-exchange was carried out in the columns by pumping 0.5 M salt solutions of the anion of interest (NaBF4 , KPF6 or LiNTf2 ) at 5 L min−1 until the retention time of toluene on the resulting columns was constant. Then the anion exchange ratio was tested by ICP.
copolymerization happened among ␥-MAPS, ViOcIm+ Br− , LMA and EDMA. Moreover, it was obvious that the matrix structure of ViOcIm+ Br− based column was much denser than the other three columns, especially ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns. As can be seen from Fig. 1b, d, f and h that the particle size of monolithic matrix increase progressively in the order ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − . of Brunauer–Emmett–Teller (BET) method was used to determine average pore diameter as 15.53 nm, 22.49 nm, 25.36 nm and 31.08 nm for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns, respectively. And the specific surface area of the monoliths was found to be 9.29 m2 g−1 , 7.95 m2 g−1 , 6.70 m2 g−1 and 6.57 m2 g−1 for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns, respectively. These results indicated that different anions had significant influence on preparation process of monolithic columns. The C%, H%, N% and O% (w/w) of these monolithic columns were determined by element analysis. As shown in Table 1, the contents of C and H immobilized on the monolithic matrix decreased in the order of ViOcIm+ Br− > ViOcIm+ BF4 − > ViOcIm+ PF6 − > ViOcIm+ NTf2 − . The content of N also decreased in the same order except for ViOcIm+ NTf2 − based column (whose anion contains N element). The cause of changes in elements content might result from the increased molecular weight of anions. Additionally, for the same anion based columns, N content of column prepared by approach 1 was lower than that of column prepared by approach 2. This may be attributed to the effect of the anions on the copolymerization process. The B%, P% and S% (w/w) of the columns prepared by approach 2 were determined as 0.63% for B, 1.75% for P and 2.14% for S. According to these results, the anion exchange ratio of the columns prepared by approach 2 could be calculated as 89.23% for BF4 − based column, 89.69% for PF6 − based column and 84.21% for NTf2 − based column. The permeability of the columns was evaluated by the Darcy’s Law [37]:
2.5. CEC procedures B0 = The resulting monolithic columns were first conditioned by running buffer for about 30 min with a manual syringe pump, and then equilibrated at a low voltage (−5 kV, ramping time for 10 min) until a stable current was obtained. Separations were performed at 25 ◦ C using different voltages and running buffer (stated in the figure captions). Stock solutions of proteins, thiourea and its analogs were prepared with distilled water (2 mg mL−1 for proteins, 0.05 mg mL−1 for thiourea and its analogs). Standard solutions of alkyl benzenes were diluted to an appropriate concentration with ethanol. All the test mixture were prepared by mixing the standard stock solutions, diluted with buffer in an appropriate ratio and stored in the refrigerator. The egg was purchased locally and egg white was diluted with 10 mM PBS (pH 7.4) in a 1:20 ratio [31]. The mixture was oscillated for 1 min and then centrifuged at 600 × g for 15 min. Prior to use, all solutions and samples were ultrasonically degassed and filtered through a 0.45 m syringe filter.
315
FL r 2 P
(1)
where F is the flow rate of the mobile phase (m3 s−1 ), is the viscosity of the mobile phase (Pa s), L is the effective length of column (m), r is the inner radius of the column (m), and P is the pressure drop of the column (Pa). The B0 value of the four monolithic columns (from ViOcIm+ Br− to ViOcIm+ NTf2 − ) was calculated as 2.05, 2.70, 3.32, 4.02 × 10−13 m2 for ACN ( = 0.38 cP), which indicated the good permeability of the resulting columns. In addition, all the columns promised excellent mechanical stability without movement of the monolithic bed under a pressure of up to 30 MPa for at least 48 h. 3.2. EOF of ILs-monolithic capillary columns Since EOF is the basic requirement for driving mobile phases through a capillary column, it is an effective indicator to the evaluation of the monolithic columns calculated as the following equations: Le × Lt V × t0
3. Results and discussion
EOF =
3.1. Characterization of ILs-monolithic capillary columns
where Le , Lt , V, t0 are the effective length (cm), total length of column (cm), the applied voltage (V), the retention time of thiourea (s), respectively. Here, ILs acted as the main source of the EOF from cathode to anode. To investigate the effects of different anions on EOF, the relationship between the pH of the running buffer and the EOF of ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − and ViOcIm+ NTf2 − based columns was examined. The results indicated
The cross-section morphology of the four monolithic columns prepared by approach 1 was obtained by SEM. In Fig. 1, it could be seen that the formed monolithic matrix was attached to the inner capillary wall well. This was because of not only the successful pretreatment of capillary by ␥-MAPS but also the perfect
(2)
316
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 1. SEM images of the ILs-monolithic capillary columns prepared by approach 1 (ViOcIm+ Br− based column: (a) 600×; (b) 5000×; ViOcIm+ BF4 − based column: (c) 600×; (d) 5000×; ViOcIm+ PF6 − based column: (e) 600×; (f) 5000×; ViOcIm+ NTf2 − based column: (g) 600×; (h) 5000×).
Table 1 The C%, H%, N% and O% (w/w) of the columns. C (%)
H (%)
N (%)
O (%)
Column 1 Br− BF4 − PF6 − NTf2 −
63.36 63.34 62.36 61.29
8.62 8.61 8.38 8.22
1.84 1.60 1.23 1.40
21.01 21.48 21.65 23.19
Column 2 BF4 − PF6 − NTf2 −
63.10 61.03 58.24
8.59 8.30 7.76
1.83 1.77 2.35
20.92 20.23 22.09
that all the columns could generate strong reversed EOF (greater than 1.60 × 10−4 cm2 V−1 s−1 ) even at a high pH 12, which indicated the successful copolymerization of ILs onto the columns. Furthermore, as shown in Fig. 2, EOF of columns prepared by approach 1 was reduced obviously from ViOcIm+ Br− to ViOcIm+ NTf2 − based column at the same test pH. It might be on account of the decrease in charge number on monolithic matrix, which was just consistent with our previous supposition that the amount of imidazolium ring on the monolithic matrix decreased from ViOcIm+ Br− to ViOcIm+ NTf2 − . Note that there was almost no visible difference in EOF among the four columns prepared by approach 2, which was almost equal to EOF of ViOcIm+ Br− based column (data not shown).
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 2. Relationship between the EOF of the ILs-monolithic capillary columns prepared by approach 1 and the pH of the running buffer. CEC conditions: 30 mM phosphate buffer (H3 PO4 -Na2 HPO4 for pH 2.0–9.0, Na2 HPO4 –NaOH for 9.0–12.0) containing 30% (v/v) ACN with various pH (pH 2.0–12.0); separation voltage, − 10 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm; EOF marker, thiourea.
317
Phenol and aniline compounds were used to further demonstrate the application of these ILs-columns. Fig. 4 showed these polar compounds were separated with elution order of phenol < aniline < m-nitrophenol < o-nitroaniline < hydroquinone, and the retention time of these analytes on the columns was prolonged in the order of ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − . Additionally, for the columns prepared by approach 1, the resolution was reduced from 1.97 to 0.75 for m-nitrophenol and o-nitroaniline in the order ViOcIm+ Br− < ViOcIm+ BF4 − < ViOcIm+ PF6 − < ViOcIm+ NTf2 − of (Fig. 4A). While, for the columns prepared by approach 2, the resolution was improved from 1.97 to 2.28 for m-nitrophenol and o-nitroaniline in the same order, and the highest separation efficiency was obtained on ViOcIm+ NTf2 − based column corresponding to column efficiency of 28 000–58 000 N m−1 (Fig. 4B). These results showed that different anions based columns had different effects on the separation of these phenol and aniline compounds, and the same anion based columns prepared by two approaches also exhibited different separation performance for these analytes. Additionally, the hydrophobicity of the long chain as well as the electropositivity and aromaticity of the imidazole on the stationary phase might cooperatively participate in the separation of phenol and aniline compounds. 3.4. Separation of thiourea and its analogs
3.3. Separation of aromatic compounds The monolithic columns prepared by two approaches were tested chromatographically to study the effects of anions on the separation performance. As shown in Fig. 3, four alkylbenzenes could be baseline separated on each column with the eluting sequence of thiourea < toluene < ethylbenzene < propylbenzene < butylbenzene. Also, the retention time for alkylbenzenes was prolonged with hydrophobicity of the anions increasing (Br− < BF4 − < PF6 − < NTf2 − ). These results indicated a reversed-phase separation of these compounds on the columns. Note that, for the same anion based columns, the retention of alkylbenzenes on columns prepared by approach 1 was stronger than that on columns prepared by approach 2. It might be ascribed to the increase in density of hydrophobic functional groups on the monolithic matrix, which was calculated based on the ratio of the content of C and H to the specific surface area of the monoliths as 77.48 mg m−2 , 90.50 mg m−2 , 105.58 mg m−2 , 105.80 mg m−2 for ViOcIm+ Br− , ViOcIm+ BF4 − , ViOcIm+ PF6 − , ViOcIm+ NTf2 − based column prepared by approach 1, respectively. As to the columns prepared by approach 2, the density of hydrophobic functional groups could not be improved by anion-exchange.
In this study, the effects of different anions based columns on the separation of analogs were also investigated. The results showed that three analogs could be baseline separated on ViOcIm+ Br− based column with the elution order dimethyl sulfoxide < thiourea < N-methylthiourea, which was consistent with the hydrophobicity of the analytes (Fig. 5A). Nevertheless, they could be partially separated on the ViOcIm+ BF4 − and ViOcIm+ PF6 − based columns and almost co-eluted on the ViOcIm+ NTf2 − based column prepared by approach 1. It might be ascribed to the different matrix structures and average pore diameters of the resulting columns based on different anions. In order to verify further, the separation of thiourea and its analogs was also attempted on columns prepared by approach 2 (Fig. 5B). The results showed that the four columns exhibited prolonged retention time and improved separation efficiency (from 31 000 to 34 000 N m−1 for dimethyl sulfoxide) with hydrophobicity of the anions increasing (Br− < BF4 − < PF6 − < NTf2 − ). These results illustrated that the hydrophobicity of the long chain and the matrix structures resulted from the anions owning different characteristics (such as size, viscosity, polarity and so on) [38–40] corporately played a role in the separation of thiourea and its analogs.
Fig. 3. Separation of alkylbenzenes on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM acetic acid buffer containing 40% (v/v) ACN at pH 3.0. Other CEC conditions were same as in Fig. 2. Analytes: 1, thiourea; 2, toluene; 3, ethylbenzene; 4, propylbenzene; 5, butylbenzene.
318
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
Fig. 4. Separation of phenol and aniline compounds on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM NaH2 PO4 buffer containing 30% (v/v) at pH 4.0; separation voltage, − 10 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm. Analytes: 1, phenol; 2, aniline; 3, m-nitrophenol; 4, o-nitroaniline; 5, hydroquinone.
Fig. 5. Separation of thiourea and its analogs on the ILs-monolithic capillary columns prepared by approach 1 (A) and prepared by approach 2 (B). (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 30 mM H3 PO4 –Na2 HPO4 buffer at pH 4.0; separation voltage, − 2 kV; injection, 0.5 psi for 5 s; detection wavelength, 214 nm. Analytes: 1, dimethyl sulfoxide; 2, thiourea; 3, N-methylthiourea.
3.5. Protein separation Here, the potential ability of all the resulting monolithic columns with different anions for the separation of protein mixture was demonstrated by four standard proteins including bovine serum albumin (BSA, molecular size 5.0 × 7.0 × 7.0 nm, isoelectric point (pI) 4.8), equine myoglobin (Mb, molecular size 4.5 × 3.5 × 2.5 nm, pI 7.0), lysozyme (Lyz, molecular size 3.0 × 3.0 × 4.5 nm, pI 11.0) and cytochrome c (Cyt c, molecular size 2.6 × 3.2 × 3.0 nm, pI 10.7). As shown in Fig. 6, they could be baseline separated within 12 min on the ViOcIm+ NTf2 − based column prepared by approach 1 using 20% (v/v) ACN, 30 mM Na2 HPO4 –citric acid buffer as background electrolyte with the elution order of Cyt c < Lyz < Mb < BSA. The separation efficiency was improved greatly compared to previous report [31], which corresponded to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c). The monoliths prepared by approach 2 exhibited poor resolution (Rs < 1) for these proteins under the same separation conditions although the elution order of the four proteins accorded with that on monoliths prepared by approach 1 (data not shown). It might be attributed to the effects of different anions on formation of matrix structures. To study separation mechanism further, the effect of ACN content (v/v) in buffer on retention factors (k) of the proteins was investigated on ViOcIm+ NTf2 − based column prepared by approach 1. The k decreased with the ACN content increasing from 15% to 40%. In our experiments, with less than 20% ACN in buffer, the great k would result in lower column efficiencies of proteins. When ACN content in buffer was over 30% the protein mixture would
be eluted together. So 20% ACN in buffer was chosen for protein separation. These results concluded that the hydrophobic interaction contributed to protein separation. Here, the organic solvent content in buffer was greatly reduced compared to the previous
Fig. 6. Separation of proteins on the ILs-monolithic capillary columns prepared by approach 1. (a) ViOcIm+ Br− based column; (b) ViOcIm+ BF4 − based column; (c) ViOcIm+ PF6 − based column; (d) ViOcIm+ NTf2 − based column. CEC conditions: 20% (v/v) ACN, 30 mM Na2 HPO4 -citric acid buffer at pH 3.0; separation voltage, −10 kV; injection, 5 psi for 10 s; detection wavelength, 210 nm. Analytes: 1, cytochrome c; 2, lysozyme; 3, equine myoglobin; 4, bovine serum albumin.
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320
319
To further validate the potential of the ViOcIm+ NTf2 − based column prepared by approach 1, it was also evaluated for the separation of the egg white sample. As shown in Fig. 8, 8 proteins were positively identified, demonstrating the enhanced separation performance of our column compared with the previously report [31]. 3.6. Reproducibility of ILs-monolithic capillary columns The reproducibility of the monolithic columns prepared by approach 1 was evaluated by measuring the relative standard deviations (RSDs) of the retention time and the peak area for four alkylbenzenes. The RSDs of the retention time and the peak area for intra-column assay were in the range of 0.88–1.42% and 3.06–5.56% (n = 5), respectively. Inter-column reproducibility for the retention time and the peak area was also evaluated which were in the range of 2.90–4.86 and 4.73–7.26%, respectively. These results indicated that the present columns possessed a robust and repeatable separation performance. Fig. 7. Effect of pH for running buffer on separation of proteins on ViOcIm+ NTf2 − based column prepared by approach 1. CEC conditions: 20% (v/v) ACN, 30 mM Na2 HPO4 –citric acid buffer at different pH. Other conditions and peak identification were as in Fig. 6.
report [9,41], which was worthy mentioning in protein separation. In addition, according to the elution order, the separation of the four proteins was consistent with ion exclusion mechanism [42] where similarly charged species to the phase are repelled, while oppositely charged and neutral species are retained, since both the imidazolium rings on stationary phase and proteins were positively charged under the test CEC conditions. The buffer pH strongly affects the protein separation because it can remarkably change the net charges. Here, the effect of pH on separation of the four standard proteins was also evaluated on ViOcIm+ NTf2 − based column prepared by approach 1. As can be seen from Fig. 7 that the retention of proteins increased, and the peaks became wider with pH increasing from 2.5 to 4.0. It might be attributed to the decrease in EOF. Besides, the decrease in the net positive charges of proteins resulted in a decrease of the repulsion forces between proteins and the imidazolium ring, which could also cause retention to increase and protein adsorption, especially BSA (pI 4.8).
4. Conclusions In this study, a series of ViOcIm+ based ILs-monolithic columns with different anions (Br− , BF4 − , PF6 − and NTf2 − ) were prepared by two approaches (“one-pot” approach and “anion-exchange” approach). The results indicated that morphology and chromatography performance of the resulting columns could be tuned by changing the kind of anions. Under CEC mode, the ViOcIm+ Br− based column exhibited highest column efficiencies for the test small molecules. The ViOcIm+ NTf2 − based column possessed the strongest retention for aromatic hydrocarbons. Their potential ability for the separation of protein mixture was demonstrated by four standard proteins, which corresponded to the highest column efficiency of 479 000 N m−1 for cytochrome c (Cyt c) on ViOcIm+ NTf2 − based column. To the best of our knowledge, this is the first time to examine and discuss the anion effects on the separation performance of monolithic columns. It means that ILs expand the variability of stationary phases of chromatography, and they are of great interest in separation science. Acknowledgements The authors are grateful to financial support from the Ministry of Science and Technology of China (Project No. 2012AA101609), the National Natural Science Foundation of China (21075089), the Ministry of Science and Technology of Tianjin (Project No. 10SYSYJC28300) and the Program for Changjiang Scholars and Innovative Research Team in University (Project No. IRT1166). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
+
−
Fig. 8. Separation of the egg white sample on ViOcIm NTf2 based column prepared by approach 1. CEC conditions: mobile phase, 40 mM Na2 HPO4 –citric acid buffer at pH 4.0; applied voltage, −2 kV; injection, 5.0 psi for 10 s; detection wavelength, 210 nm. Analytes: the egg white sample.
[12] [13]
G. Guiochon, J. Chromatogr. A 1168 (2007) 101–168. J.X. He, G.Z. Fang, Q.L. Deng, S. Wang, Anal. Chim. Acta 692 (2011) 57–62. J. Rozenbrand, W.P. van Bennekom, J. Sep. Sci. 34 (2011) 1934–1944. J.N. Hay, H.M. Raval, Chem. Mater. 13 (2011) 3396–3403. S.D. Chambers, F. Svec, J.M.J. Frechet, J. Chromatogr. A 1218 (2011) 2546–2552. Z. Lin, H. Huang, X. Sun, Y. Lin, L. Zhang, G. Chen, J. Chromatogr. A 1246 (2012) 90–97. J. Liu, L. Ren, Y. Liu, H. Li, Z. Liu, J. Chromatogr. A 1228 (2012) 276–282. M. Iwasaki, N. Sugiyama, N. Tanaka, Y. Ishihama, J. Chromatogr. A 1228 (2012) 292–297. D. Bandilla, C.D. Skinner, J. Chromatogr. A 1004 (2003) 167–179. H.S. Schrekker, M.P. Stracke, C.M.L. Schrekker, J. Dupont, Ind. Eng. Chem. Res. 46 (2007) 7389–7392. Z.B. Alfassi, R.E. Huie, B.L. Milman, P. Neta, Anal. Bioanal. Chem. 377 (2003) 159–164. J.L. Anthony, E.J. Maginn, J.F. Brennecke, J. Phys. Chem. B 105 (2001) 10942–10949. X.A. Tian, X.P. Zhang, L. Wei, S.J. Zeng, L. Huang, S.J. Zhang, Green Chem. 12 (2010) 1263–1273.
320 [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
C.-C. Liu et al. / Analytica Chimica Acta 804 (2013) 313–320 P. Domínguez de María, Angew. Chem. 120 (2008) 7066–7075. I.J.B. Lin, C.S. Vasam, J. Organomet. Chem. 690 (2005) 3498–3512. L.H. Lu, W. Chen, Adv. Mater. 22 (2010) 3745–3748. C.M.S.S. Neves, J.F.O. Granjo, M.G. Freire, A. Robertson, N.M.C. Oliveira, J.A.P. Coutinho, Green Chem. 13 (2011) 1517–1526. L. Vidal, M.L. Riekkola, A. Canals, Anal. Chim. Acta 715 (2012) 19–41. R. Giernoth, Angew. Chem. Int. Ed. 49 (2010) 2834–2839. T. Payagala, Y. Zhang, E. Wanigasekara, K. Huang, Z.S. Breitbach, P.S. Sharma, L.M. Sidisky, D.W. Armstrong, Anal. Chem. 81 (2009) 160–173. T.T. Ho, C.Y. Chen, Z.G. Li, T.C.C. Yang, M.R. Lee, Anal. Chim. Acta 712 (2012) 72–77. W.C. Siegler, J.A. Crank, D.W. Armstrong, R.E. Synovec, J. Chromatogr. A 1217 (2010) 3144–3149. C.F. Poole, S.K. Poole, J. Sep. Sci. 34 (2011) 888–900. J.L. Anderson, D.W. Armstrong, Anal. Chem. 77 (2005) 6453–6462. A. Martin-Calero, V. Pino, J.H. Ayala, V. González, A.M. Afonso, Talanta 79 (2009) 590–597. M. Bian, Z.J. Zhang, H. Yin, J. Pharm. Biomed. Anal. 58 (2012) 163–167. Y. Tian, R. Feng, L.P. Liao, H.L. Liu, H. Chen, Z.R. Zeng, Electrophoresis 29 (2008) 3153–3159. H.F. Han, J. Li, X.S. Wang, X. Liu, S.X. Jiang, J. Sep. Sci. 34 (2011) 2323–2328.
[29] J. Li, H.F. Han, Q. Wang, X. Liu, S.X. Jiang, J. Sep. Sci. 34 (2011) 1555–1560. [30] H. Han, Q. Wang, X. Liu, S. Jiang, J. Chromatogr. A 1246 (2012) 9–14. [31] D. Zhou, L.N. Xiang, R.J. Zeng, F.H. Cao, X.X. Zhu, Y.M. Wang, J. Sep. Sci. 34 (2011) 3441–3450. [32] M.H. Kleemiss, M. Gilges, G. Schomburg, Electrophoresis 14 (1993) 515–522. [33] C.A. Lucy, A.M. MacDonald, M.D. Gulcev, J. Chromatogr. A 1184 (2008) 81–105. [34] Y. Wang, Q.L. Deng, G.Z. Fang, M.F. Pan, Y. Yu, S. Wang, Anal. Chim. Acta 712 (2012) 1–8. [35] Y.N. Hsieh, W.Y. Ho, R.S. Horng, P.C. Huang, C.Y. Hsu, H.H. Huang, C.H. Kuei, Chromatographia 66 (2007) 607–611. [36] R. Marcilla, J.A. Blazquez, J. Rodriguez, J.A. Pomposo, D. Mecerreyes, J. Polym. Sci. Part A: Polym. Chem. 42 (2004) 208–212. [37] R.D. Stanelle, L.C. Sander, R.K. Marcus, J. Chromatogr. A 1100 (2005) 68–75. [38] R. Losada, C. Wandrey, Macromolecules 42 (2009) 3285–3293. [39] Y.H. Shih, B. Singco, W.L. Liu, C.H. Hsu, H.Y. Huang, Green Chem. 13 (2011) 296–299. [40] H. Tokuda, K. Hayamizu, K. Ishii, M. Abu Bin, Susan Hasan, M. Watanabe, J. Phys. Chem. B 108 (2004) 16593–16600. [41] M.H. Wu, R.A. Wu, R.B. Li, H.Q. Qin, J. Dong, Z.B. Zhang, H.F. Zou, Anal. Chem. 82 (2010) 5447–5454. [42] P.R. Haddad, M. Noviˇc, J. Chromatogr. A 1118 (2006) 19–28.